How to install Trafodion on existing and running CDH5.0 Cluster? Without affecting existing settings

You should be able to start from step 10.

https://wiki.trafodion.org/wiki/index.php/Trafodion_Installation

Could you please tell me what version of the installer you have?

Thanks!
Sent from my iPhone

> On Aug 7, 2014, at 8:27 AM, "Mohan" <email address hidden> wrote:
>
> New question #252766 on Trafodion:
> https://answers.launchpad.net/trafodion/+question/252766
>
> We are planning to install Trafodion on already existing cluster. How to do We do it.
>
> The document says following instructions "https://wiki.trafodion.org/wiki/index.php/Installing_Cloudera".
>
> In Document never says how to install on existing CDH cluster. Always you are going to talk about fresh installation of CDH Trafodion.
>
> Please provide document how to install Trafodion on Existing running cluster?
>
> Please provide it ASAP
>
> Regards
> Mohan
>
> --
> You received this question notification because you are an answer
> contact for Trafodion.

On Cloudera installation page, there is a link at the bottom of the page "installing Trafodion" (Step#23)

Please follow the instructions to install Trafodion on existing running cluster:
https://wiki.trafodion.org/wiki/index.php/Trafodion_Installation#STEP_START_INSTALL

Please make sure to follow steps under "Preparing Your Cluster Environment" before you start installing Trafodion.

I get below error when I start from Step 10.

trafodion_server-0.8.3.tgz
installer-0.8.3_v2.tar.gz

[root@xxx xxx]# ./trafodion_mods --trafodion_build /opt/cloudera/trafodion_downloads/trafodion_server-0.8.3.tgz --userid admin --password xxx --url http://hdpcmt01.cps.intel.com:7180
cat: /opt/cloudera/trafodion_downloads/installer/../node_list: No such file or directory
cat: /opt/cloudera/trafodion_downloads/installer/../my_nodes: No such file or directory

***ERROR: Please check input parameters

You will need to execute trafodion_setup script (step#7) prior to executing trafodion_installer script.
Please follow steps under "Preparing Your Cluster Environment" and start following instructions from step#1.

If start from Step 7 ...

If you are installing MapR, run the ./trafodion_setup script by providing the following parameters:

Again it will go and install Cloudera manager.

My question is How to install Trafodion on existing and running CDH5.0 Cluster?

Step No7 .. What it will do , If Already Cloudera manager is up and running ?

trafodion Supoort Cloudera 5.0.1 Version?

The trafodion_setup script does preliminary setup (creating trafodion user, configuring environment to support trafodion_installer script, etc). It does not install Cloudera manager. Once steps under "Preparing Your Cluster Environment" are followed and all *.tar.gz files are in place (step#1 to 5), skip step#6 (as Cloudera is already up and running) and do the followings:

- Execute trafodion_setup script (step#7 to 9)
- Execute trafodion_mods script (step#10)
- Execute trafodion_installer script (step#11 to 13)

Can you please let us know which version of HBase you have?
At this moment Trafodion supports HBase 0.94 and you can soon expect it supporting 0.96 and 0.98?

I have already completed steps 1 to 5 Successfully.

But, In step 7. If you are installing MapR, run the ./trafodion_setup script by providing the following parameters:

Does it apply only for MapR ? I understood the step 7 needs to be done only on MapR not on Cloudera

You can use trafodion_setup regardless of the Hadoop distribution you are using.

Sorry for the confusion.

--Amanda

-----Original Message-----
From: <email address hidden> [mailto:<email address hidden>] On Behalf Of Mohan
Sent: Thursday, August 07, 2014 8:13 PM
To: Moran, Amanda
Subject: Re: [Question #252766]: How to install Trafodion on existing and running CDH5.0 Cluster? Without affecting existing settings

Question #252766 on Trafodion changed:
https://answers.launchpad.net/trafodion/+question/252766

    Status: Answered => Open

Mohan is still having a problem:
I have already completed steps 1 to 5 Successfully.

But, In step 7. If you are installing MapR, run the ./trafodion_setup script by providing the following parameters:

Does it apply only for MapR ? I understood the step 7 needs to be done only on MapR not on Cloudera

--
You received this question notification because you are an answer contact for Trafodion.

We have already installed Cloudera manager and other components like (HDFS,MR,HBASE,ZOOKEEPER etc) . This step 7 will again installs Cloudera manager?

trafodion_setup will not reinstall Cloudera.

--Amanda

-----Original Message-----
From: <email address hidden> [mailto:<email address hidden>] On Behalf Of Mohan
Sent: Thursday, August 07, 2014 9:02 PM
To: Moran, Amanda
Subject: Re: [Question #252766]: How to install Trafodion on existing and running CDH5.0 Cluster? Without affecting existing settings

Question #252766 on Trafodion changed:
https://answers.launchpad.net/trafodion/+question/252766

    Status: Answered => Open

Mohan is still having a problem:
We have already installed Cloudera manager and other components like (HDFS,MR,HBASE,ZOOKEEPER etc) . This step 7 will again installs Cloudera manager?

--
You received this question notification because you are an answer contact for Trafodion.

Hi When I ran step 7.. I got below error

./trafodion_setup --nodes "hdpcdn02 hdpcdn03 hdpcdn04" --home_dir /opt/cloudera/trafodion --clustername mycldtrafodion

su: user trafodion does not exist
su: user trafodion does not exist
su: user trafodion does not exist
su: user trafodion does not exist
chmod: cannot access `/opt/cloudera/trafodion/trafodion/.ssh/config': No such file or directory
***INFO: creating .qpidports file
su: user trafodion does not exist
***INFO: creating .bashrc file
cp: cannot create regular file `/opt/cloudera/trafodion/trafodion/.bashrc': No such file or directory
chown: invalid user: `trafodion.trafodion'
***INFO: creating sqconfig file
cp: cannot create regular file `/opt/cloudera/trafodion/trafodion/sqconfig': No such file or directory
chown: invalid user: `trafodion.trafodion'
***INFO: Setting up userid trafodion on all other nodes in cluster
cp: cannot stat `/opt/cloudera/trafodion/trafodion': No such file or directory
chown: cannot access `/opt/cloudera/trafodion/trafodion_downloads/installer/../trafodion': No such file or directory
pdcp@hdpcmt01: can't stat /opt/cloudera/trafodion/trafodion_downloads/installer/../trafodion
hdpcdn04: cp: cannot stat `/opt/cloudera/trafodion/trafodion_downloads/installer/../trafodion': No such file or directory
pdsh@hdpcmt01: hdpcdn04: ssh exited with exit code 1
hdpcdn02: cp: cannot stat `/opt/cloudera/trafodion/trafodion_downloads/installer/../trafodion': No such file or directory
pdsh@hdpcmt01: hdpcdn02: ssh exited with exit code 1
hdpcdn03: cp: cannot stat `/opt/cloudera/trafodion/trafodion_downloads/installer/../trafodion': No such file or directory
pdsh@hdpcmt01: hdpcdn03: ssh exited with exit code 1
***INFO: Creating known_hosts file for all nodes
su: user trafodion does not exist
***ERROR: Unable to ssh to node hdpcdn02
***ERROR: Unable to create Trafodion userid: trafodion

I am running step number 7 from different node not on these nodes "hdpcdn02 hdpcdn03 hdpcdn04" I am running from Mgmt node which has got ssh access to all. So the step number 7 uses this script also "traf_add_user" So this got limitations. I think we need to run the step 7 from first node of "hdpcdn02 hdpcdn03 hdpcdn04"

I am able to complete Step 7 successfully after including the node name which I was running.

But Step 8 through Below error

 Certificate/Private key created in directory :/opt/cloudera/trafodion/trafodion/sqcert
***********************************************************

***********************************************************
 Updating Authentication Configuration
***********************************************************
Creating folders for storing certificates

Copying the log4j and log4cpp configuration files
Copying the log4j and log4cpp configuration files
***INFO: copying /opt/cloudera/trafodion/trafodion/sqcert directory to all nodes
***INFO: Start of DCS install
***INFO: untarring build file /opt/cloudera/trafodion/trafodion/dcs_build.tgz
***INFO: modifying /opt/cloudera/trafodion/trafodion/trafodion/dcs-0.8.3/conf/dcs-env.sh
***INFO: modifying /opt/cloudera/trafodion/trafodion/trafodion/dcs-0.8.3/conf/dcs-site.xml
***INFO: creating /opt/cloudera/trafodion/trafodion/trafodion/dcs-0.8.3/conf/servers file
***INFO: End of DCS install.
***INFO: copying install to all nodes
***INFO: starting Trafodion instance
Checking orphan processes.
Removing old mpijob* files from /opt/cloudera/trafodion/trafodion/trafodion/tmp

Removing old monitor.port* files from /opt/cloudera/trafodion/trafodion/trafodion/tmp

Executing sqipcrm (output to sqipcrm.out)
Starting the SQ Environment (Executing /opt/cloudera/trafodion/trafodion/trafodion/sql/scripts/gomon.cold)
Background SQ Startup job (pid: 29634)

# of SQ processes: 18 ..................................
Error while executing the startup script!!!
Checking if processes are up.
Checking attempt: 1; user specified max: 1. Execution time in seconds: 0.

The SQ environment is not up at all. Check the logs.

Process Configured Actual Down
------- ---------- ------ ----
DTM 4 0 \$tm0 \$tm1 \$tm2 \$tm3

Please check the SQ shell log file : /opt/cloudera/trafodion/trafodion/trafodion/logs/sqmon.log

SQ Startup (from /opt/cloudera/trafodion/trafodion/trafodion/sql/scripts) Failed

./opt/cloudera/trafodion/trafodion/trafodion/dcs-0.8.3/bin/start-dcs.sh found.
Starting the DCS environment now
starting master, logging to /opt/cloudera/trafodion/trafodion/trafodion/dcs-0.8.3/bin/../logs/dcs-trafodion-1-master-hdpcdn02.out
hdpcdn02: starting server, logging to /opt/cloudera/trafodion/trafodion/trafodion/dcs-0.8.3/bin/../logs/dcs-trafodion-1-server-hdpcdn02.out
hdpcdn03: starting server, logging to /opt/cloudera/trafodion/trafodion/trafodion/dcs-0.8.3/bin/../logs/dcs-trafodion-2-server-hdpcdn03.out

You can monitor the SQ shell log file : /opt/cloudera/trafodion/trafodion/trafodion/logs/sqmon.log

Startup time 0 hour(s) 10 minute(s) 49 second(s)
***ERROR: sqstart failed with RC=1. Check /opt/cloudera/trafodion/trafodion/trafodion/sqmon.log file for details.

Adding more errors from a file " /opt/cloudera/trafodion/trafodion/trafodion/logs/sqmon.log"

[root@hdpcdn02 ~]# cat /opt/cloudera/trafodion/trafodion/trafodion/logs/sqmon.log
Processing cluster.conf on local host hdpcdn02
[SHELL] Shell/shell Version 1.0.1 Release 0.8.3 (Build release [0.8.3-0-g0187afb_Bld32], date 20140722_173451)

[SHELL] %
! Start the monitor processes across the cluster
startup
[SHELL] %startup
[SHELL] - Warning using shell.env

[$Z010P6Z] %
exit
[$Z010P6Z] %exit
Able to connect to the SQ monitor.
Continuing with the Startup...

Processing cluster.conf on local host hdpcdn02
[$Z010PAF] Shell/shell Version 1.0.1 Release 0.8.3 (Build release [0.8.3-0-g0187afb_Bld32], date 20140722_173451)

[$Z010PAF] %
set CLUSTERNAME=mycldtrafodion
[$Z010PAF] %set CLUSTERNAME=mycldtrafodion

[$Z010PAF] Configuration Change Notice for Group: CLUSTER Key: CLUSTERNAME
[$Z010PAF] %
set SQ_MBTYPE=64
[$Z010PAF] %set SQ_MBTYPE=64

[$Z010PAF] Configuration Change Notice for Group: CLUSTER Key: SQ_MBTYPE
[$Z010PAF] %
set MY_NODES= -w hdpcdn02 -w hdpcdn03 -w hdpcdn04 -w hdpcmt01
[$Z010PAF] %set MY_NODES= -w hdpcdn02 -w hdpcdn03 -w hdpcdn04 -w hdpcmt01
[$Z010PAF] Configuration Change Notice for Group: CLUSTER Key: MY_NODES

[$Z010PAF] %
exit
[$Z010PAF] %exit
SQ_START_SEAPILOT: 0
Processing cluster.conf on local host hdpcdn02
[$Z010PAI] Shell/shell Version 1.0.1 Release 0.8.3 (Build release [0.8.3-0-g0187afb_Bld32], date 20140722_173451)

[$Z010PAI] %
set SQ_SEAPILOT_SUSPENDED=1
[$Z010PAI] %set SQ_SEAPILOT_SUSPENDED=1

[$Z010PAI] %
[$Z010PAI] Configuration Change Notice for Group: CLUSTER Key: SQ_SEAPILOT_SUSPENDED
exit
[$Z010PAI] %exit
Processing cluster.conf on local host hdpcdn02
[$Z010PAL] Shell/shell Version 1.0.1 Release 0.8.3 (Build release [0.8.3-0-g0187afb_Bld32], date 20140722_173451)

[$Z010PAL] %
! Start DTM
set DTM_RUN_MODE=2
[$Z010PAL] %set DTM_RUN_MODE=2
s[$Z010PAL] Configuration Change Notice for Group: CLUSTER Key: DTM_RUN_MODE
et SQ_AUDITSVC_READY=1
[$Z010PAL] %set SQ_AUDITSVC_READY=1
s[$Z010PAL] Configuration Change Notice for Group: CLUSTER Key: SQ_AUDITSVC_READY
et DTM_TLOG_PER_TM=1
[$Z010PAL] %set DTM_TLOG_PER_TM=1
[$Z010PAL] Configuration Change Notice for Group: CLUSTER Key: DTM_TLOG_PER_TM

[$Z010PAL] %
exit
[$Z010PAL] %exit
Processing cluster.conf on local host hdpcdn02
[$Z010PAW] Shell/shell Version 1.0.1 Release 0.8.3 (Build release [0.8.3-0-g0187afb_Bld32], date 20140722_173451)
set {process $tm0} TMASE=TLOG0
[$Z010PAW] %set {process $tm0} TMASE=TLOG0
exec {type dtm, nowait, name $tm0, nid 0, out stdout_dtm_0} tm
[$Z010PAW] %exec {type dtm, nowait, name $tm0, nid 0, out stdout_dtm_0} tm
[$Z010PAW] NewProcess failed to spawn, error=Error in spawn call
[$Z010PAW] Process $TM0 terminated normally. Nid=0, Pid=31786
delay 5
[$Z010PAW] %delay 5
exit
[$Z010PAW] %exit
Processing cluster.conf on local host hdpcdn02
[$Z010PGZ] Shell/shell Version 1.0.1 Release 0.8.3 (Build release [0.8.3-0-g0187afb_Bld32], date 20140722_173451)
set {process $tm1} TMASE=TLOG1
[$Z010PGZ] %set {process $tm1} TMASE=TLOG1
exec {type dtm, nowait, name $tm1, nid 1, out stdout_dtm_1} tm
[$Z010PGZ] %exec {type dtm, nowait, name $tm1, nid 1, out stdout_dtm_1} tm
[$Z010PGZ] NewProcess failed to spawn, error=Error in spawn call
exit
[$Z010PGZ] %exit
[$Z010PGZ] Process $TM1 terminated normally. Nid=1, Pid=29996
Processing cluster.conf on local host hdpcdn02
[$Z010PHS] Shell/shell Version 1.0.1 Release 0.8.3 (Build release [0.8.3-0-g0187afb_Bld32], date 20140722_173451)
set {process $tm2} TMASE=TLOG2
[$Z010PHS] %set {process $tm2} TMASE=TLOG2
exec {type dtm, nowait, name $tm2, nid 2, out stdout_dtm_2} tm
[$Z010PHS] %exec {type dtm, nowait, name $tm2, nid 2, out stdout_dtm_2} tm
[$Z010PHS] NewProcess failed to spawn, error=Error in spawn call
[$Z010PHS] Process $TM2 terminated normally. Nid=2, Pid=28446
exit
[$Z010PHS] %exit
Processing cluster.conf on local host hdpcdn02
[$Z010PI3] Shell/shell Version 1.0.1 Release 0.8.3 (Build release [0.8.3-0-g0187afb_Bld32], date 20140722_173451)
set {process $tm3} TMASE=TLOG3
[$Z010PI3] %set {process $tm3} TMASE=TLOG3
exec {type dtm, nowait, name $tm3, nid 3, out stdout_dtm_3} tm
[$Z010PI3] %exec {type dtm, nowait, name $tm3, nid 3, out stdout_dtm_3} tm
[$Z010PI3] NewProcess failed to spawn, error=Error in spawn call
[$Z010PI3] Process $TM3 terminated normally. Nid=3, Pid=26083
delay 5
[$Z010PI3] %delay 5
exit
[$Z010PI3] %exit
Checking if processes are up.
Checking attempt: 60; user specified max: 60. Execution time in seconds: 592.

Process Configured Actual Down
------- ---------- ------ ----
DTM 4 0 \$tm0 \$tm1 \$tm2 \$tm3

The dtm process(es) are Not Ready yet. Stopping further startup (if any).
Error while executing the startup script!!!

Please check the SQ shell log file : /opt/cloudera/trafodion/trafodion/trafodion/logs/sqmon.log

SQ Startup (from /opt/cloudera/trafodion/trafodion/trafodion/sql/scripts) Failed

Yes, this scripts expects to be ran on the first node.

--Amanda

-----Original Message-----
From: <email address hidden> [mailto:<email address hidden>] On Behalf Of Mohan
Sent: Friday, August 08, 2014 12:57 AM
To: Moran, Amanda
Subject: Re: [Question #252766]: How to install Trafodion on existing and running CDH5.0 Cluster? Without affecting existing settings

Question #252766 on Trafodion changed:
https://answers.launchpad.net/trafodion/+question/252766

Mohan gave more information on the question:
I am running step number 7 from different node not on these nodes
"hdpcdn02 hdpcdn03 hdpcdn04" I am running from Mgmt node which has got ssh access to all. So the step number 7 uses this script also "traf_add_user" So this got limitations. I think we need to run the step 7 from first node of "hdpcdn02 hdpcdn03 hdpcdn04"

--
You received this question notification because you are an answer contact for Trafodion.